Identification of an Intestinal Folate
Transporter and the Molecular Basis
for Hereditary Folate Malabsorption
Andong Qiu,1,2Michaela Jansen,3Antoinette Sakaris,4Sang Hee Min,1Shrikanta Chattopadhyay,1
Eugenia Tsai,1,2Claudio Sandoval,5Rongbao Zhao,1,2Myles H. Akabas,1,3and I. David Goldman1,2,*
1Department of Medicine
2Department of Molecular Pharmacology
3Department of Physiology and Biophysics
4Department of Obstetrics and Gynecology
Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461, USA
5Department of Pediatrics, New York Medical College, Valhalla, NY 10595, USA
Folates are essential nutrients that are required
cesses. While folates are absorbed in the acidic
milieu of the upper small intestine, the under-
lying absorption mechanism has not been
defined. We now report the identification of
a human proton-coupled, high-affinity folate
transporter that recapitulates properties of
folate transport and absorption in intestine and
in various cell types at low pH. We demonstrate
that a loss-of-function mutation in this gene is
the molecular basis for hereditary folate malab-
sorption inafamilywiththis disease. This trans-
porter was previously reported to be a lower-
affinity, pH-independent heme carrier protein,
HCP1. However, the current study establishes
that a major function of this gene product is
proton-coupled folate transport required for
folate homeostasis in man, and we have thus
amended the name to PCFT/HCP1.
Folates are essential cofactors that are required for the
provision of one-carbon moieties in key biosynthetic and
epigenetic processes (Stover, 2004). Folate deficiency is
prevalent in underdeveloped countries, and even in the
Western world, subtle deficiency is a public health prob-
lem that is most notable in its association with neural tube
defects in the developing embryo (Eichholzer et al., 2006).
Mammals cannot synthesize folates; hence, dietary sour-
ces must meet metabolic needs, necessitating anefficient
intestinal absorptive mechanism. Absorption of folates
occurs primarily in the duodenum and upper jejunum
and involves a carrier-mediated process with a low-pH
optimum that operates efficiently within the acidic micro-
climate of the intestinal surface in this region (Selhub and
Rosenberg, 1981; Mason and Rosenberg, 1994; McEwan
et al., 1990). The specificity and other properties of this
process have been well established, and similar folate
transport activities with a low-pH optimum have been
identified inother normal tissuesand in human solid tumor
cell lines (Horne, 1993; Zhao et al., 2004). Despite the pre-
valence and importance of this process, a folate transport
protein with a low-pH optimum has not been identified.
There are two known highly specific mammalian folate
transporters. Their properties were the subject of a recent
review (Matherly and Goldman, 2003). The reduced folate
carrier(SLC19A1) isafacilitative transporter withthechar-
acteristics of an anion exchanger. There are two GPI-
linked folate receptors, high-affinity binding proteins that
mediate cellular uptake by an endocytic mechanism. Fo-
late receptor expression in small intestine is negligible.
While the reduced folate carrier is expressed on the brush
border membrane of intestinal cells, this transporter has
a neutral pH optimum and a specificity profile that differs
substantially from that observed in intestinal folate ab-
sorption and transport into intestinal cells and cells of
other tissue origin at low pH (Selhub and Rosenberg,
1981; Mason and Rosenberg, 1994; Wang et al., 2004).
Further, when reduced folate carrier function is lost due
to deletion, mutation, or loss of expression of the gene,
the low-pH folate transport activity remains intact (Zhao
et al., 2004, 2005b; Wang et al., 2005).
This report describes the identification of a proton-
coupled, electrogenic, high-affinity folate transporter
with properties that are similar to folate transport in intes-
tinal and other cells at low pH. A database mining ap-
proach was utilized based on the conserved amino acid
sequence of SLC19 family members and the screening
of candidate mRNAs in cell lines developed in this labora-
Cell 127, 917–928, December 1, 2006 ª2006 Elsevier Inc. 917
low pH activity was either retained or markedly decreased
(Zhao et al., 2004, 2005a). Having identified this carrier as
a candidate intestinal folate transporter, we demonstrate
a loss-of-function mutation in this gene in a family with
the syndrome of hereditary folate malabsorption.
Identification of a Low-pH Folate Transporter
To identify the low-pH folate transporter, the Ensembl
human peptide database was mined at low stringency as
described in the Experimental Procedures. Twenty-three
functions were identified, and mRNA expression levels
were screened intwo HeLacell linesdeveloped inthislab-
oratory: (1) the HeLa-R5 (Zhao et al., 2004) cell line, which
has a genomic deletion of the reduced folate carrier gene
but a high level of low-pH folate transport activity and (2)
the HeLa-R1 line (Zhao et al., 2005a), a HeLa-R5 deriva-
tive, cloned by antifolate selective pressure, in which the
low-pH activity is markedly diminished. Gene 21 (to be
referred to as G21) was identified in this screen as a likely
candidate based on a high mRNA level in HeLa-R5 cells
versus avery lowlevelof expression inHeLa-R1 cells(Fig-
ure 1A). G21 (GenBank accession number NP_542400) is
predicted to be a membrane protein of 459 amino acids
with a MW of z50 kDa. A BLAST search of the Swissprot
database revealed that the human protein shares 91%
similarity and 87%identity to both its mouse and rat coun-
terparts (GenBank accession numbers AAH57976 and
AAH89868). During the course of the studies described
below, this protein was reported by another group to be
a heme carrier protein (HCP1) and entered into GenBank
as such (Shayeghi et al., 2005). This protein was desig-
nated as SLC46A1 in the Human Genome Organization
(HUGO) Nomenclature Committee Database.
Folate transport properties mediated by this carrier
were assessed by injection of G21 cRNA into Xenopus
laevis oocytes. As indicated in Figure 1B, uptake of 2 mM
[3H]folic acid and [3H]methotrexate (MTX) was increased
Figure 1. Identification and Initial Characterization of the Low-pH Folate Transporter
(A) G21 mRNA levels inHeLa,HeLa-R5,and HeLa-R1cells determined byquantitative RT-PCR; G21 mRNA inHeLa cells wasassigned thevalueof 1.
The values are the mean ± SEM for two independent experiments. (B) Functional expression of G21 in Xenopus oocytes. [3H]MTX or [3H]folic acid
(2 mM) uptake was assayed at pH 5.5 over 30 min. (C) Initial uptake of [3H]MTX or [3H]folic acid (0.5 mM), at pH 5.5 and 37?C, into HepG2 cells stably
transfected with pcDNA3.1(+) (Mock-HepG2) or pcDNA3.1(+)G21 (G21-HepG2). (D) Initial uptake of [3H]MTX or [3H]folic acid (0.5 mM), at pH 5.5 and
37?C, into HeLa cells transiently transfected with pcDNA3.1(+) (Mock-HeLa) or pcDNA3.1(+)G21 (G21-HeLa). The data in (B)–(D) are the mean ± SEM
from three independent experiments. (E) Detection of G21 protein expressed in Xenopus oocytes and HepG2 cells by SDS-PAGE and western blot-
The blot is representative of three independent experiments. (F) Plasma membrane localization of G21 protein in HeLa cells transiently transfected
with G21 cDNA detected by immunofluorescence. The green fluorescence shows the localization of G21 protein, and the red fluorescence indicates
the counterstaining by propidium iodide. The image shown is representative of three independent studies.
918 Cell 127, 917–928, December 1, 2006 ª2006 Elsevier Inc.
>200-fold in G21 cRNA-injected oocytes at pH 5.5 as
compared to water-injected oocytes. Similarly, uptake of
these folates into HepG2 cells stably transfected with
G21 cDNA (Figure 1C) and into HeLa cells transiently
transfected with G21 cDNA (Figure 1D) was increased
>30- and >13-fold, respectively. A western blot using
a polyclonal antibody directed to the C terminus of G21
indicated a broad band in G21 cRNA-injected, but not
water-injected, Xenopus oocytes and in HepG2 cells sta-
bly transfected with this cDNA, but not in mock-trans-
fected cells (Figure 1E). Differences in migration in the
two systems may be due to differences in glycosylation.
When expressed in HeLa cells, G21 protein targeted to
the plasma membrane in permeabilized cells as demon-
strated with the polyclonal antibody (Figure 1F). Staining
could not be detected in mock-transfected HeLa cells
(data not shown).
The pH Dependence of G21-Mediated Transport
Transport of folates mediated by G21 was highly pH de-
pendent, as illustrated for tritiated folic acid (Figure 2A),
(6S)5-methyltetrahydrofolate [(6S)5-MTHF, Figure 2B],
MTX (Figure S1A in the Supplemental Data available with
FTHF; Figure S1B]. Activity was highest at the lowest pH
and declined as the pH was increased, although the pat-
tern of decrease, and the extent of retention of activity at
neutral pH, was different among the folates. For both
(6S)5-MTHF, the major blood folate in man and rodents,
and (6S)5-FTHF, there was residual activity at pH 7.5,
and for all the folates there was substantial activity at
pH 6.5. The dependence of G21 activity on the inward-
directed electrochemical H+gradient was further charac-
terized by exposure of Xenopus oocytes to carbonyl
cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP), an
ionophore that collapses the transmembrane H+gradient
(Benz and McLaughlin, 1983). As shown in Figure 2C and
Figure S1C, respectively, at an extracellular pH of 5.5,
[3H]folic acid and [3H]MTX uptake were markedly de-
creased by 10 mM FCCP.
The Kinetics of Folate Transport Mediated by G21
as a Function of pH
Uptake mediated by G21 conformed to Michaelis-Menten
kinetics with a Km for [3H]folic acid uptake, which in-
creased as pH increased, from 1.3 ± 0.1 mM at pH 5.5 to
56.2 ± 5.6 mM at pH 7.5 (Figures 2D–2H). Vmaxand Km
were evaluated as a function of pH using the same batch
of oocytes, each injected with the same amount of G21
cRNA. As shown in Figure 2G, the Vmaxfor folic acid
uptake decreased from 13 pmol/oocyte/hr at pH 5.5 to
5 pmol/oocyte/hr at pH 7.5. The major change (2.4-fold)
occurred over the pH range of 5.5 to 6.5, with only a small
decline beyond pH 6.5. The pattern of change was differ-
ent for the folic acid Kmmeasured at the same time. There
was a small (2.6-fold) increase in Kmas the pH was in-
creased from 5.5 to 7.0, but there was a 18-fold increase
when the pH was increased from 7.0 to 7.5. Figure 2H
indicates a 2- to 3-fold higher uptake Kmfor MTX as
compared to folic acid over this pH range, with a similar
pattern of increase as the pH was increased.
Structural Specificity of G21-Mediated Transport
Figure 3A shows the inhibitory effects of a variety of com-
pounds, at a concentration of 20 mM, on uptake of 2 mM
[3H]folic acid into Xenopus oocytes injected with G21
cRNA. There was a high degree of structural specificity
among the folate/antifolate compounds tested; peme-
trexed was the most potent competitor. This is an antifo-
late inhibitor of thymidylate synthase with the highest af-
finity for the low-pH transporter in human tumor cell lines
specific; the natural 6S isomer of 5-FTHF was more
potent than the unnatural 6R isomer in reducing [3H]folic
acid uptake. The 6S isomer of 5-MTHF had effects com-
parable to those of (6S)5-FTHF. MTX was a less effective
transport Kms. PT523 is a dihydrofolate reductase inhibi-
tor with a high affinity for the reduced folate carrier at
high and low pH but a very low affinity (Ki> 50 mM) for the
low-pH folate transporter in HeLa and other cells (Zhao
et al., 2005b; Wang et al., 2004). This antifolate did not in-
hibit under these conditions. All the folates that inhibited
[3H]folic acid uptake are transport substrates of G21.
Thus, the inhibition of [3H]folic acid uptake was due to
competition fortransport viaG21.Figure3Bcomparesthe
structures of these folate compounds. The major differ-
ence between PT523 and the other folates/antifolates is
at the g-carboxyl moiety, suggesting the importance of
this group to folate binding to G21. Bromosulphopthalein,
para-aminohippuric acid, taurocholic acid, cholic acid,
and estrone-3-sulfate, substrates for organic anion solute
carriers (SLC21 and SLC 22) (Hagenbuch and Meier,
2003; Koepsell and Endou, 2004), did not inhibit [3H]folic
acid influx. Hemin was a weak inhibitor. At a [3H]folic
acid concentration of 2 mM, 100 mM hemin inhibited up-
take by 42% ± 7% in Xenopus oocytes injected with
G21 cRNA and by 30% ± 5% in HepG2 cells stably trans-
fected with G21 cDNA. In the same experiments, 100 mM
nonlabeled folic acid inhibited 2 mM [3H]folic acid uptake
by 92% ± 2% and 90% ± 0.02%, respectively (based on
the average of three separate experiments at pH 6.5).
Electrophysiological Properties of G21-Mediated
Transport in Xenopus Oocytes
Electrophysiological characteristics were evaluated in
two-electrode voltage-clamp experiments. In G21 cRNA-
injected oocytes, folic acid, (6S)5-MTHF, and MTX in-
duced currents of up to 80 nA at a ?80 mV holding poten-
tial (Figure 4A). These folates did not induce current in
water-injected oocytes. These substrate-induced cur-
rents imply that net charge translocation occurred across
the cell membrane during each transport cycle. Consis-
tent with this, substrate currents were proportional to both
applied voltage and substrate concentration, increasing
Cell 127, 917–928, December 1, 2006 ª2006 Elsevier Inc. 919
with more negative voltage and higher substrate concen-
tration (Figures 4B–4D and Figure 5).
In order to distinguish between whether folate transport
was coupled to proton transport (i.e., protons are trans-
ported with the folate) or whether protons just bind to
the transporter and regulate its activity, we determined
the effect of changing extracellular pH from 5.5 to 7.5
on the substrate current-voltage relationship. The reversal
potential of the current-voltage relationship is the voltage
at which the substrate-induced current is zero (the x axis
intercept in Figures 4C and 4D) and is a measure of the
net driving force for substrate transport. If folate and
Figure 2. pH Dependence and Kinetics of G21-Mediated Uptake of Tritiated Folates in Xenopus Oocytes
(A and B) Uptake of 2 mM [3H]folic acid (A) and [3H](6S)5-MTHF (B) in water- and G21 cRNA-injected oocytes over 30 min. (C) Water- and G21 cRNA-
injected oocytes were preincubated at pH 5.5 with a series of FCCP concentrations for 20 min. Uptake of [3H]folic acid (2 mM) was subsequently
assessed at pH 5.5 over 1 hr. (A)–(C) are representative of three independent studies. (D–F) Uptake of [3H]folic acid over 60 min at pH 5.5 (D),
6.5 (E), and 7.5 (F) as a function of the extracellular folic acid concentration, [Folic acid]e. The lines were generated, and kinetic constants were
calculated, based on Michaelis-Menten kinetics (V = Vmax[S]/(Km+ [S])). The data are representative of two to four experiments as summarized in
(H). (G) Effects of extracellular pH on [3H]folic acid uptake Kmand Vmax. All measurements were made in a single batch of oocytes with the injection
of the same amount of G21 cRNA. (H) A summary of uptake Kmfor MTX and folic acid as a function of extracellular pH. All data are the mean ± SEM.
920 Cell 127, 917–928, December 1, 2006 ª2006 Elsevier Inc.
proton transport are coupled, then the reversal potential
would become more negative as the extracellular pH is
raised, and the slope would decrease. In contrast, if the
folate transport rate were only regulated by pH, then the
slope of the current-voltage relationship would decrease,
but the reversal potential would not change. As the pH in-
creased, the reversal potential became more negative
(Figures 4C and 4D). Changing the pH from 5.5 to 6.5
shifted the reversal potentials for MTX and (6S)5-MTHF
by ?8 mV and ?6 mV, respectively, whereas increasing
the pH from 6.5 to 7.5 shifted the reversal potentials by
?36 mV and ?30 mV, respectively. This change in the re-
versal potential with a change in pH provides direct evi-
dence that the transmembrane proton gradient is coupled
to folate transport. These results are consistent with the
loss of net proton influx driving transport, as the extracel-
pH of?7.3.At pH 7.5, the transmembrane proton gradient
is close to zero, and folate influx is driven solely by the
folate concentration gradient due to the higher extracel-
lular folate concentration. Thus, the reversal potential is
more negative at pH 7.5 than at lower pHs where the
transmembrane proton gradient also contributes to the
net driving force for transport.
Figure 5A shows the currents recorded from an individ-
ual oocyte as the extracellular MTX concentration was
Figure 3. Substrate Specificity of G21 in Xenopus Oocytes
(A) Uptake of 2 mM [3H]folic acid was assessed at pH 5.5 over 30 min
in the absence (control) or presence of 20 mM nonlabeled folates, anti-
folates, or other organic anions. The data are the mean ± SEM of two
(B) The structures of folate and antifolate compounds studied.
Figure 4. Electrophysiological Characterization of the G21
Transporter in Xenopus Oocytes
(A) Substrate-induced currents recorded by two-electrode voltage-
clamp from G21 expressing oocytes at a ?80 mV holding potential
with concentrations of folates 20–25 times the Kms at pH 5.5. Currents
for all substrates were measured in individual oocytes (n = 8 oocytes).
A representative experiment from one oocyte is shown.
(B) Currents from a G21-expressing oocyte, left superfused with buffer
at pH 5.5, right superfused with MTX at 25 times the Kmat pH 5.5. Re-
sponses to depolarizing and hyperpolarizing voltage-clamp steps are
2 s in 10 mV increments from ?100 mV to +30 mV. The dashed line
indicates the level of I = 0.
(C and D) Current-voltage relationships as a function of extracellular pH
for MTX (C) and (6S)5-MTHF (D) with concentrations of 20–25 times the
Cell 127, 917–928, December 1, 2006 ª2006 Elsevier Inc. 921
increased from 0.1 to 120 mM. Current was detected at a
MTX concentration as low as 0.1 mM. Figure 5B illustrates
the dependence of current on the MTX concentration at
pH 5.5 and 6.5. The Kms for folic acid, (6S)5-MTHF, and
MTX (Figure 5C), and their pH dependence, were compa-
rable in the electrophysiological and the tritiated substrate
uptake assays (Figures 2H and 5C). The Kms at pH 6.5 for
folic acid and MTX were nearly four times greater than
those at pH 5.5, while the Kmfor (6S)5-MTHF was only
minimally increased. The relative current magnitude in-
duced by the different substrates was assessed by apply-
ing saturating concentrations of each substrate to the
same oocyte at pH 5.5 (Figure 4A). When normalized to
the (6S)5-MTHF-induced current magnitude, the current
amplitudes were 39% ± 6% and 86% ± 11% (n = 8) larger
for folic acid and MTX, respectively, despite the fact that
the latter substrates had higher Kms (Figure 5C). This im-
plies that the transport Vmaxfor these substrates was
higher than that for (6S)5-MTHF.
An attempt was made to determine if hemin transport
could be detected by current flow into oocytes injected
with G21 cRNA. In two separate experiments using two
different batches of oocytes, three to four oocytes for
each condition, folic acid produced substantial currents,
while no current could be detected with 100 mM hemin
either when hemin was stabilized with 0.1% bovine serum
albumin or with 200 mM arginine (data not shown). If hemin
were transported by G21 in an electrogenic fashion similar
to the folates, then the ‘‘expected’’ currents from this ap-
proximately EC50hemin concentration (reported hemin
Kmof 125 mM [Shayeghi et al., 2005]) would be well within
the detection limits of this system.
Lack of Impact of Other Extracellular Ions
on G21-Mediated Transport into Oocytes
Substitution of extracellular Na+with N-methyl-glucos-
amine did not decrease [3H]MTX uptake into G21 cRNA-
Similarly, folic acid-induced currents were unchanged
when K+, Ca2+, or Mg2+was removed from the extracel-
lular solution or when the extracellular Cl?concentration
was reduced from 95.6 mM to 5.6 mM by replacement
of NaCl with Na-gluconate. Thus, folate transport was
not dependent on extracellular Na+, K+, Ca2+, Mg2+, or
Cl?, implying that none of these ions are involved in the
folate transport cycle (data not shown).
G21 Expression in Human Tissues and Tumor
G21mRNA levels were examinedin a variety ofhuman tis-
sues by northern blotting (Figure 6A). After hybridization
weredetected withmolecular sizesofz2.7 kband2.1kb.
The latter, short form was dominant, with a molecular size
consistent with G21 mRNA in the NCBI database (2.096
amounts of both G21 mRNA forms were detected in kid-
ney, liver, placenta, small intestine, and spleen, and to a
lesser extent, in colon and testis. There was very low ex-
pression in brain, lung, stomach, heart, and muscle. A
and kidney, and an even smaller transcript was detected
in liver (?0.5 kb). Quantitative RT-PCR was employed to
further quantify G21 mRNA expression in intestine along
with two human tumor cell lines (Figure 6B). Expression
in Caco2 cells, a colon carcinoma cell line, was >7-fold
higher than thatin HeLacells. In intestine, the highest level
of mRNA expression was in duodenum, with lesser ex-
pression in jejunum and lower levels in ileum, cecum,
segments of the colon, and rectum. Consistent with the
northern blot, there was a high level of expression in liver.
Figure 5. Concentration Dependence of Substrate-Induced
(A) Currents recorded from an individual oocyte in response to (from
left to right) 0.1, 1, 3, 4, 5, 7, 10, 20, 40, and 120 mM MTX (pH 5.5,
Vh= ?80 mV). Bars above current traces denote time of substrate
(B) Currents obtained as described in (A) at pH 5.5, and in similar
experiments at pH 6.5, were normalized to the maximum current Imax
for each oocyte and plotted as a function of substrate concentration.
The normalized currents from different experiments were fit to the
equation I = (Imax3[S])/(Km+[S]) to obtain Km. Data are the mean ± SEM
for three to eight oocytes from two toads.
(C) Summary of the electrophysiologically determined Kmvalues for
folic acid, (6S)5-MTHF, and MTX as a function of extracellular pH.
The data are the mean ± SEM for three to seven experiments.
922 Cell 127, 917–928, December 1, 2006 ª2006 Elsevier Inc.
The Impact of Suppression of G21 Expression
by Interfering RNA on Folate Transport in Human
As indicated above, Caco2 cells have a very high level of
G21 mRNA expression. To establish the extent to which
constitutive low-pH folate transport activity in Caco2 cells
could be attributed to this transporter, two small hairpin
RNA (shRNA) vectors, targeted to two different regions of
the G21 transcript, were stably cotransfected into Caco2
cells.This resultedina 55%reductionin[3H]folicacidinflux
at pH 5.5 and a similar (50%) decrease in G21 mRNA as
quantified by RT-PCR (Figure 6C). Wild-type Caco2 cells
were subjected to transient transfection with small interfer-
ing RNA (siRNA) duplex using the Amaxa system, resulting
in a 60% reduction in [3H]MTX influx and a 50% decrease
in G21 mRNA as compared to negative siRNA-transfected
cells (Figure 6D). When the stably transfected Caco2 cells
sulting in an 80% decrease in [3H]MTX influx at pH 5.5 and
a 75% decreasein G21 mRNA ascomparedto vector con-
trol-transfected cells (Figure 6E). Taken together, these
studies demonstrate that G21 is the major, and possibly
the only, low-pH folate transporter in Caco2 cells.
Figure 6. G21 mRNA Levels in Human Tissues and Tumor Cell Lines and Suppression of G21 Expression and Folate Transport
Activity in Caco2 Cells by Interfering RNA
(A) Expression of G21 mRNA levels in human tissues by northern blotting. b-actin was the loading control. Open and filled triangles indicate the
location of two G21 major mRNA transcripts. (B) G21 mRNA levels in two human cell lines and tissues were determined by quantitative RT-PCR.
G3PDH mRNA was the house-keeping gene to normalize G21 expression. The ordinate represents expression of G21 mRNA relative to expression
inHeLa cells assigned thevalueof 1.(C) Theimpactof stable transfection of G21 shRNA constructs into Caco2 cells on G21 mRNA levels determined
by RT-PCR (inset) and uptake of [3H]folic acid (0.5 mM) at pH 5.5 and 37?C for 2 min as compared to cells transfected with negative shRNA plasmids.
(D) The impact of transient transfection of Caco2 cells with siRNA duplex, using the Amaxa system, on G21 mRNA (inset) and uptake of [3H]MTX
(0.5 mM) at pH 5.5 and 37?C for 2 min as compared to cells with scrambled (Neg) siRNA. (E) Caco2 cells stably transfected with the G21 shRNA con-
structs were subjected to transient transfection with siRNA duplex usingthe Amaxa system, and both G21 mRNA levels (inset) and uptakeof [3H]MTX
(0.5 mM) at pH 5.5 and 37?C for 2 min were assessed. The data in (B)–(E) are the mean ± SEM from three independent experiments.
Cell 127, 917–928, December 1, 2006 ª2006 Elsevier Inc. 923
An Analysis of the Role of G21 in the Pathogenesis
of Hereditary Folate Malabsorption in a Family
with This Disease
Hereditary folate malabsorption (OMIM 229050) is a rare
recessive familial disorder characterized by signs and
symptoms of folate deficiency that appear within a few
months after birth. Infants exhibit low blood and cerebro-
spinal fluid folate levels with anemia, diarrhea, immune
deficiency, infections, and neurological deficits. There is
a profound defect in intestinal folate absorption (Geller
et al., 2002). To determine whether an alteration in G21
is the molecular basis for this disorder, blood was ob-
tained from a family with progeny manifesting this disease
that were the subject of a recent report (Geller et al., 2002;
Figure 7A). The mother and father were both normal, one
of their children died in infancy, and two of their other
daughters displayed classical signs and symptoms of
hereditary folate malabsorption. One daughter was diag-
nosed at the age of 8 months with a serum folate level of
0.2 nM (normal: 10–30 nM), and the other daughter was
diagnosed at the age of 2 months with a blood folate level
of less than 0.2 nM. Both children were treated with high
doses oforal 5-FTHF withcompleteresolution ofthe signs
and symptoms of their disease; they have developed nor-
mally and remain completely well on their folate supple-
ment now at ages 9 and 6.
G21 is composed of five exons and four introns
(Figure 7B). Each of the five exons of G21 along with their
flanking intron regions was sequenced from these family
members. This revealed a homozygous mutation in G21
from both daughters and the same mutation in one allele
of G21 from each parent (Figure 7C). This G to A mutation
(position 5882; GenBank accession number DQ496103) is
located in the splice acceptor of intron 2 (intron 2/exon 3
boundary). To determine the consequences of this geno-
mic mutation on RNA splicing, the exon 3 region of G21
mRNA was analyzed by RT-PCR from these family mem-
bers. Two DNA fragments of 579 bp and 495 bp were de-
tected from the parents’ transformed lymphocyte cDNAs,
whereas in the daughters only a single DNA fragment was
detected with a size identical to the shorter fragment from
the parents. In comparison, the control cDNA from normal
intestine exhibited a single amplified DNA fragment that
was identical in size to the longer DNA fragment from
the parents (Figure 7D). Subsequent DNA sequencing
showed that the longer DNA fragment contained exon 3,
Figure 7. Genetic and Functional Analy-
sis of G21 in a Family with Hereditary Fo-
late Malabsorption Syndrome
(A) Pedigree of a family with hereditary folate
malabsorption. P1 and P2 are the parents; D1
and D2 are the affected daughters.
(B) Genomic organization of G21 and splicing
of the wild-type and mutated G21 mRNA.
(C) Representative chromatograms of se-
quenced DNA showing a heterozygous muta-
tion in the father (P1) and a homozygous muta-
tion in a daughter (D1) in G21.
(D) Agarose gelanalysisofRT-PCRproducts of
mutated and wild-type G21 cDNA from family
members. The control is a PCR fragment de-
rived from normal human intestinal cDNA.
HeLa cells. Exon 3-deleted (G21/-Exon 3),
wild-type G21 cDNA (G21), and empty plasmid
vector (Mock) were transiently transfected into
HeLa cells. Uptake of (6S)[3H]5-MTHF (0.5 mM)
was examined at pH 5.5 and 37?C over 2 min.
The data are the mean ± SEM for four indepen-
(F) Western blot analysis of wild-type (lane 2)
and exon 3-deleted (lane 3) G21 proteins in
transiently transfected HeLa cells. Empty plas-
mid (Mock) was transfected into HeLa cells as
a control (lane 1). b-actin was the loading con-
trol. Theblotisrepresentativeof threeindepen-
(G) Subcellular localization of wild-type (G21)
expressed in transiently transfected HeLa cells
as determined by immunofluorescence. The
image shown is representative of three inde-
924 Cell 127, 917–928, December 1, 2006 ª2006 Elsevier Inc.
whereas the shorter one did not. Hence, the single-nucle-
otide mutation of G21 results in skipping of exon 3 and
consequent in-frame deletion of 28 amino acids. Interest-
the GenBank (accession number BC010691) and appears
to represent an alternatively spliced form (Figure 7B).
When transiently transfected into HeLa cells, this mutated
of (6S)5-MTHF uptake (Figure 7E). Western blot analysis
showed that the mutated G21 protein was less efficiently
expressed and had a lower molecular weight than the
wild-type protein when transiently transfected into HeLa
cells (Figure 7F). Further, immunofluorescence analysis
indicated that, when expressed in HeLa cells, the mutated
G21 carrier was trapped intracellularly without detectable
localization to the cell membrane (Figure 7G). Thus, the
parents carried both a functional wild-type and the non-
functional mutated G21 mRNA, and the daughters with
the disease had only the nonfunctional mutated G21
mRNA. No mutation was detected in the reduced folate
carrier mRNA amplified from the transformed lymphocyte
cDNA from both daughters.
These studies have identified G21, previously identified
as HCP1 (SLC46A1), as a proton-coupled, electrogenic
folate transporter that has the properties of the low-pH
in intestinal and other human cells—a high affinity for folic
acid (Ki?0.6 mM) and a low affinity for the PT523 antifolate
(Ki? >50 mM) at pH 5.5. This is in contrast to what is
observed for the reduced folate carrier, a facilitative trans-
porter (SLC19A1) ubiquitously expressed in human tis-
sues. This carrier has low affinity for folic acid (Ki?200
mM), high affinity for PT523 (Ki?0.7 mM), and a pH opti-
mum of 7.4. The affinity of the reduced folate carrier for
these folates, also in contrast to the low-pH transporter,
does not change appreciably between pH 5.5 and 7.4
(Matherly and Goldman, 2003; Wang et al., 2004).
The identification of a loss-of-function mutation in hcp1
that results in the deletion of the third exon in a family with
hereditary folate malabsorption establishes that this gene
is an intestinal transporter required for normal folate ab-
sorption and homeostasis. Accordingly, we amend the
name of the transporter to PCFT/HCP1. This takes into
tissues and may have functions beyond intestinal folate
absorption. Consistent with the role of PCFT/HCP1 in
intestinal absorption are the following observations: (1)
PCFT/HCP1 mRNA is expressed in small intestine, partic-
ularly in the duodenum and to a lesser extent in jejunum,
segments that account for the bulk of folate absorption.
These are areas in which the pH at the microenvironment
of the intestinal surface is in the range of 6.0–6.2 (McEwan
et al., 1990). (2) PCFT/HCP1 protein is localized to the api-
cal brush border of intestinal cells (Shayeghi et al., 2005).
(3) PCFT/HCP1 is highly expressed in Caco2 cells, which
manifest a high level of low-pH folate transport activity
and have been used as a model for intestinal transport
(Hidalgo et al., 1989). (4) This constitutive folate transport
activity in Caco2 cells can be nearly abolished (?80%
suppression) by PCFT/HCP1 interfering RNA. The pH
dependence of folate transport mediated by PCFT/HCP1
is consistent with studies in everted jejunal sacs and rings
(Mason and Rosenberg, 1994). Quantitatively, in rat jeju-
num brush border membranes the uptake Kmfor folic acid
increased from 0.6 mM at pH 5.5 to 12.3 mM at pH 7.4 and
was competitively inhibited (Ki = 1.4 mM) by racemic
5-MTHF (Mason et al., 1990; Selhub and Rosenberg,
1981). The identification of this carrier not only confirms
the earlier conclusion that low-pH folate transport must
be mediated by a mechanism genetically distinct from
the reduced folate carrier, but also argues against an im-
absorption, as has been proposed (Said, 2004). Hence,
while the reduced folate carrier is expressed in the upper
small intestine, its activity must be negligible since it can-
not compensate for the loss of the PCFT/HCP1 in individ-
uals with hereditary folate malabsorption under the acidic
conditions of the absorptive surface. Likewise, the level of
reduced folate carrier expression and function in the more
alkaline distal small intestinal compartments is, appar-
ently, insufficient to meet folate requirements at usual
dietary folate levels. However, with the pharmacologic
doses of 5-FTHF that are used to treat individuals with
this disease (Geller et al., 2002), this mechanism may be
the route of delivery under these conditions.
This transporter was recently reported to be an intesti-
nal heme carrier protein (HCP1). The murine ortholog
was characterized as pH independent over a range of
6.5 to 8.0 and with a Kmof 125 mM for [55Fe]hemin uptake
in HeLa cells infected with a cDNA-containing adenovirus
of transport activity was observed in Xenopus oocytes mi-
croinjected withthemurinecRNA.Transporter mRNA was
highly expressed in duodenum, and protein was localized
to the apical brush border membrane of murine intestinal
take into both Xenopus oocytes injected with PCFT/HCP1
cRNA and HepG2 cells stably transfected with this carrier.
However, we were unable to detect a hemin-induced cur-
rent in Xenopus oocytes expressing this transporter under
conditions in which currents for the folate compounds
were easily detectable. This implies that either there is
no electrogenic hemin transport or that the Vmaxfor hemin
transport must be more than an order of magnitude lower
than that of folic acid. Based on the high affinity of this
transporter for folates (approximately two orders of mag-
nitude greater than the reported affinity for [55Fe]hemin),
the high degree of specificity (including stereospecificity)
for the folate/antifolate compounds, and the etiologic
itary folate malabsorption, it is clear that folates are major
physiological substrates for this transporter. Further,
the apparent complete correction of the hematological
Cell 127, 917–928, December 1, 2006 ª2006 Elsevier Inc. 925
disorder with high doses of folates in individuals with he-
reditary folate malabsorption who lack both wild-type
copies of this gene argues against an important role of
this carrier in the intestinal absorption of iron (Geller
et al., 2002).
While PCFT/HCP1 operates most optimally at low pH,
there is residual transport activity for 5-MTHF, the major
blood folate (Opladen et al., 2006), at pH 7.4, suggesting
that PCFT/HCP1 plays a role in the delivery of this folate
to systemic cells under physiological conditions. Hence,
the physiological importance of PCFT/HCP1 may extend
to other organs in which PCFT/HCP1 mRNA is expressed,
especially where transport activity at low pH has been
documented, i.e., liver, which is a major folate storage
site (Horne, 1993), but an acidic microenvironment is not
present. From a pharmacological perspective, PCFT/
HCP1 may play an important role in the delivery of antifo-
lates into the acidic interior of solid tumors (Helmlinger
et al., 1997; Wike-Hooley et al., 1984). The data in this
paper, along with previous reports (Zhao et al., 2005b;
Wang et al., 2004), suggest that transport of pemetrexed,
a new-generation antifolate now in clinical use, would be
ity for PCFT/HCP1 at acidic and neutral pH.
Besides its role in cellular transport, the PCFT/HCP1
may contribute to folate receptor-mediated endocytosis
(Anderson et al., 1992). In this process, folate binds to
glycosyl-phosphoinositol (GPI)-linked folate receptors at
the cell surface, which are internalized in endocytic vesi-
cles. Within the cytoplasm, the vesicles acidify, resulting
in a marked transvesicular proton gradient. Acidification
results in the dissociation of folate from the receptor and
a strong driving force that would favor folate export from
the vesicle via the PCFT/HCP1 (Murphy et al., 1984; Pau-
los et al., 2004). Similarly, PAT1-mediated export of amino
acids from lysosomal vesicles in brain neurons has been
proposed in addition to the role of this proton-coupled
transporter in intestinal amino acid absorption (Boll
et al., 2002; Sagne et al., 2001).
The identification of a molecular basis underlying folate
transport mediated by a proton-coupled carrier offers
a new dimension to the understanding of the physiology
of folate transport, in particular, intestinal folate absorp-
tion and the mechanism of delivery of folates to peripheral
tissues in which this activity is expressed. The molecular
basis for hereditary folate malabsorption has been estab-
lished. It is now possible to assess the role that alterations
in this transporter might play in folate deficiency condi-
tions. The observation that patients with this disease
have no evidence of neural tube defects and that neuro-
logical deficits and other signs and symptoms appear
months after birth implies that this gene is not absolutely
required for delivery of folates to cells in the neural crest
during embryonic neural tube formation. Rather, polymor-
phisms or mutations in this gene might contribute to
maternal folate deficiency, especially in the developing
world, compounding dietary folate deficiency and thereby
increasing thechances of neural tubedefects inthe devel-
oping embryo (Eichholzer et al., 2006). Indeed, the inci-
dence of hereditary folate malabsorption may be greater
than previously appreciated, since most infants with this
disorder in areas with endemic folate deficiency would
be expected to die early in infancy, undiagnosed.
Cell Lines and Cell Culture Conditions
HeLa, HepG2, and Caco2 cells were obtained from the American Type
Tissue Collection (Manassas, Va). HeLa, HeLa-R5, and HepG2 cells
were maintained in RPMI 1640 medium. HeLa-R1 cells were main-
tained in the same medium at pH 6.9 in the presence of 500 nM
MTX. Caco2 cells were grown in DMEM. All media were supplemented
with 10% fetal bovine serum (Gemini Bio-Products, Calabasas, CA),
2 mM glutamine, 20 mM 2-mercaptoethanol, 100 units/ml penicillin,
and 100 mg/ml streptomycin.
[3H]folic acid, [3H]MTX, [3H](6S)5-FTHF, and [3H](6S)5-MTHF were ob-
tained from Moravek Biochemicals (Brea, CA) and purity monitored
and maintained by HPLC. (6S)5- and (6R)5-FTHF and (6S)5-MTHF
were obtained from Schircks Laboratories (Jona, Switzerland). PT523,
an antifolate analog, was a gift from Andre Rosowsky (Dana-Farber
Cancer Institute, Boston, MA). Folic acid, MTX, FCCP, hemin, estrone-
3-sulfate, taurocholic acid, cholic acid, sulfobromophthalein, and
para-amino hippurate were obtained from Sigma-Aldrich (St. Louis,
MO). Hemin was dissolved in DMSO as a 5 mM stock solution.
FCCP was dissolved in 95% ethanol to a concentration of 5 mM.
Database Mining of the Human Genome
The Ensembl human peptide database was blasted with the search
parameter of Distant Homology to identify distant homologues using
the conserved domains across species of the three SLC19 family
members (GenBankaccession number pfam01770.12) and the human
reduced folate carrier (GenBank accession number NP_919231) as
query. The predicted proteins, with similarity to SLC19 family trans-
porters and unknown function, were chosen and used for subsequent
screening of differential mRNA expression between HeLa-R5 and
HeLa-R1 cells by RT-PCR.
Cloning and Construction of G21
The open reading frame of G21 was amplified from cDNA of HeLa-R5
cells with pfuUltra DNA polymerase (Stratagene, Cedar Creek, TX) and
primers that contain BglII restriction sites (underlined in Table S1) and
subsequently cloned into the BglII site of the pSPT64 vector for syn-
thesis of capped sense G21 cRNA from the SP6 promoter using the
mMESSAGE mMACHINE system (Ambion, Austin, TX), and into the
BamHI site of pcDNA3.1(+) to generate pcDNA3.1(+)G21.
Construction of G21 shRNA
The Silencer Express (Human U6) kit (Ambion, Austin, TX) was used
according to the manufacturer’s protocol to produce shRNA expres-
sion cassettes (SECs), which were screened by transient transfection
into HeLa cells followed by measurement of MTX initial uptake and
quantitative RT-PCR of G21 mRNA. The most effective SEC targeting
G21 mRNA (1000-ACTAATCGGCTATGGTTCT-1020; GenBank ac-
cession number NM_080669) and the negative SEC were cloned into
the pSEC hygromycin vector (Ambion). A commercial shRNA targeting
G21 mRNA (841-CGATCCATTGTCCAGCTCTAT-861) and a negative
nonsilencing shRNA in a pSM2 retroviral vector were obtained from
Open Biosystems (Huntsville, AL).
Transfection of plasmid DNA was performed in HepG2, HeLa, and
Caco2 cells with Lipofectamine 2000 (Invitrogen). HepG2 cells, stably
926 Cell 127, 917–928, December 1, 2006 ª2006 Elsevier Inc.
transfected with either pcDNA3.1(+) or pcDNA3.1(+)G21, were gener-
ated by G418 selection (600 mg/ml). Double selection with puromycin
(5 mg/ml) and hygromycin (50 mg/ml) was adopted to obtain stably
transfected Caco2 cells with both G21-silencing shRNA vectors, or
with both nonsilencing negative control plasmids.
Amaxa Nucleofection of G21 siRNA Oligonucleotides
The Nucleofector II unit and the Nucleofector cell line kit T (Amaxa Inc.,
Gaitherburg, MD) were employed to nucleofect Caco2 cells with
SMARTpool siRNA containing four different siRNA duplexes (catalog
#L-018653, Dharmacon, Inc., Lafayette, CO) that target G21 mRNA
or siCONTROL nontargeting siRNAs (catalog #D-001210-01, Dharma-
con), which lack homology to any human gene. The nucleofected cells
were assayed on day 3 postseeding for initial [3H]MTX uptake and G21
mRNA expression by quantitative RT-PCR.
Uptake Studies in Xenopus Oocytes
Defolliculated Xenopus laevis oocytes were prepared as described
(Jansen and Akabas, 2006) and injected with 50 nl of water or G21
cRNA (30 ng). Radiotracer uptake was determined 3 or 4 days later.
Seven to ten oocytes were incubated in 500 ml of modified Barth’s
solution [MBS; 88 mM NaCl, 2.4 mM NaHCO3, 2.5 mM Na pyruvate,
1 mM KCl, 0.82 mM MgSO4, 0.41 mM CaCl2, 0.3 mM Ca(NO3)2,
15 mM MES or HEPES], and uptake of tritiated folate substrates was
assessed at room temperature. Uptake was halted by the addition of
ice-cold MBS (pH 7.5). Oocytes were washed ten times thereafter
and solubilized with 10% SDS for measurement of radioactivity. To
collapse the pH gradient across the oocyte membrane, seven to ten
oocytes were incubated in MBS (pH 5.5) containing 0, 10, 20, 40, or
60 mM FCCP for 20 min, and uptake of transport substrates was
assessed at pH 5.5.
Transport of Folates in HepG2, HeLa, and Caco2 Cells
Initial uptake of tritiated folates in HepG2, HeLa, or Caco2 cells was
assessed using a protocol designed for rapid uptake determinations
in cells growing in monolayer culture in liquid scintillation vials (Sharif
and Goldman, 2000), except that cells were incubated at pH 7.4 and
37?C for 20 min before initiation of uptake. Substrate uptake was nor-
malized to protein content.
Electrophysiological Analyses in Xenopus Oocytes
2.5 mM KCl, 1 mM MgCl2, 2.3 mM CaCl2, 5 mM HEPES, 5% horse
serum [pH 7.5]). Electrophysiological recordings were conducted 3–7
days after cRNA injection in buffer (90 mM NaCl, 1 mM KCl, 1 mM
(Jansen and Akabas, 2006). Oocyte holding potential was ?80 mV
for Km determination. For current-voltage (I-V) relationships, from
a ?60 mV holding potential step changes in membrane potential
were applied for 2 s in 10 mV increments between ?100 and 30 mV
in the absence and presence of substrate.
Production of Peptide Antibody and Immunofluorescence
To generate antisera to human G21 protein, a peptide ([C]ADPH
LEFQQFPQSP) corresponding to amino acids 446–459 of this protein
was synthesized, conjugated with KLH, and injected into rabbits by
Open Biosystems. The IgG fraction was isolated from the antiserum
using a protein A-conjugated agarose column (Bio-Rad, Hercules,
CA), and antibodies specific for G21 were purified with the Sulfolink
Trial Kit (Pierce, Rockford, IL). Immunofluorescence was performed
using affinity-purified anti-G21 and FITC-conjugated swine anti-rabbit
antibody (Dako, Carpinteria, CA). HeLa cells were permeabilized with
0.2% Triton X-100 in phosphate buffer (PBS) at pH 7.4 for 15 min.
The stained samples were mounted on slides with Vectashield mount-
ing medium containing 1.5 mg/ml propidium iodide (Vecta Laborato-
ries, Burlingame, CA).
SDS-PAGE and Western Blotting
Water- and G21 cRNA-injected oocytes were homogenized in MBS
with a protease inhibitor cocktail (Sigma-Aldrich). The homogenate
was spun at 1000 3 g and 4?C for 5 min to collect supernatant, and
the membrane fraction was pelleted by centrifugation at 13,200 3 g
and 4?C for 30 min and resuspended in MBS with protease inhibitors.
To obtain HepG2 cell membranes, cells were incubated on ice for
taining protease inhibitors, following whichthemembranefractionwas
pelleted by centrifugation at 13,200 3 g and 4?C for 10 min and resus-
pended in the same buffer. SDS-PAGE and protein blotting were con-
ducted to detect G21 protein using rabbit anti-G21 antibody and sec-
ondary goat anti-rabbit IgG-horseradish peroxidase conjugate (Cell
Signaling Technology, Danvers, MA).
A human PolyA+northern RNA blot containing polyA+RNA (2 mg per
lane) of 12 tissues (Origene, Rockville, MD) was hybridized with32P-
dCTP-labeled cDNA probes from a G21 cDNA segment (97–396 bp;
GenBank accession number NM_080669) overnight at 42?C followed
by four 20 min high-stringency washes at 65?C. b-actin mRNA was
probed as the loading control.
cDNA was synthesized from DNase I-treated total RNA from HeLa,
HeLa-R5, HeLa-R1, and Caco2 cells with Superscript Reverse Tran-
scriptase II (Invitrogen). cDNAs of the human digestive system were
obtained from Clontech (Mountain View, CA). Real-time PCR was per-
formed with SYBR green PCR Master Mix (Applied Biosystems, Foster
City, CA) and primers specific for G21 (Table S1). G3PDH or b-actin
was simultaneously amplified with specific primers (Table S1) as
housekeeping genes to normalize the G21 expression.
Analysis of G21 in a Family with Hereditary Folate
Members of a family with hereditary folate malabsorption were studied
according to a protocol approved by the Albert Einstein College of
Medicine IRB (CCI #2006-279), and informed consent in the subjects’
native language was obtained. Whole blood was used for isolation of
genomic DNA by a Genomic DNA Purification Kit (Gentra Systems,
Minneapolis, MN) and to generate EBV-transformed human B-lym-
phoblastoid cell lines in the Einstein Human Genetics Cell Culture
Core. Each G21 exon with flanking regions was amplified using Taq
in Table S2. G21 or RFC cDNA was amplified from lymphoblastoid
cells by RT-PCR. PCR products were gel purified and sequenced in
an ABI 3730 DNA Analyzer (Applied Biosystems). The mutated region
was verified by sequencing both DNA strands. An expression vector of
the mutated G21 in which exon 3 was skipped (GenBank accession
number BC010691) was purchased from Open Biosystems and, along
with pcDNA3.1(+)G21 (wild-type), was used for an assay of transport
function. Western blot analysis on whole-cell lysate and cellular local-
ization were performed as described above.
The Supplemental Data include one supplemental figure and two sup-
plemental tables and can be found with this article online at http://
This work was supported by grants CA82621 (I.D.G.) and GM77660
(M.H.A.) from the National Institutes of Health. We thank Dr. Alan Fin-
kelstein for helpful discussions.
Cell 127, 917–928, December 1, 2006 ª2006 Elsevier Inc. 927
Received: August 18, 2006 Download full-text
Revised: September 19, 2006
Accepted: September 27, 2006
Published: November 30, 2006
Anderson, R.G.W., Kamen, B.A., Rothberg, K.G., and Lacey, S.W.
(1992). Potocytosis: Sequestration and transport of small molecules
by caveolae. Science 255, 410–411.
Benz, R., and McLaughlin, S. (1983). The molecular mechanism of
action of the proton ionophore FCCP (carbonylcyanide p-trifluoro-
methoxyphenylhydrazone). Biophys. J. 41, 381–398.
Boll, M., Foltz, M., Rubio-Aliaga, I., Kottra, G., and Daniel, H. (2002).
Functional characterization of two novel mammalian electrogenic
proton-dependent amino acid cotransporters. J. Biol. Chem. 277,
Eichholzer, M., Tonz, O., and Zimmermann, R. (2006). Folic acid:
A public-health challenge. Lancet 367, 1352–1361.
Geller, J., Kronn, D., Jayabose, S., and Sandoval, C. (2002).Hereditary
folate malabsorption: Family report and review of the literature. Medi-
cine (Baltimore) 81, 51–68.
Hagenbuch, B., and Meier, P.J. (2003). The superfamily of organic
anion transporting polypeptides. Biochim. Biophys. Acta 1609, 1–18.
Helmlinger, G., Yuan, F., Dellian, M., and Jain, R.K. (1997). Interstitial
ments reveal a lack of correlation. Nat. Med. 3, 177–182.
Hidalgo, I.J., Raub, T.J., and Borchardt, R.T. (1989). Characterization
of the human colon carcinoma cell line (Caco-2) as a model system
for intestinal epithelial permeability. Gastroenterology 96, 736–749.
Horne, D.W. (1993). Transport of folates and antifolates in liver. Proc.
Soc. Exp. Biol. Med. 202, 385–391.
Jansen, M., and Akabas, M.H. (2006). State-dependent cross-linking
of the M2 and M3 segments: Functional basis for the alignment of
GABAA and acetylcholine receptor M3 segments. J. Neurosci. 26,
Koepsell,H.,and Endou, H.(2004).The SLC22 drug transporter family.
Pflugers Arch. 447, 666–676.
Mason, J.B., and Rosenberg, I.H. (1994). Intestinal absorption of
folate. In Physiology of the Gastrointestinal Tract, L.R. Johnson, ed.
(New York: Raven Press), pp. 1979–1995.
Mason, J.B., Shoda, R., Haskell, M., Selhub, J., and Rosenberg, I.H.
transport inthesmall intestine. Biochim. Biophys.Acta 1024,331–335.
Matherly, L.H., and Goldman, D.I. (2003). Membrane transport of
folates. Vitam. Horm. 66, 403–456.
McEwan, G.T., Lucas, M.L., Denvir, M., Raj, M., McColl, K.E., Russell,
R.I.,and Mathan,V.I.(1990).Acombined TDDA-PVCpHandreference
electrode foruseintheuppersmall intestine. J.Med.Eng.Technol.14,
Murphy, R.F., Powers, S., and Cantor, C.R. (1984). Endosome pH
measured in single cells by dual fluorescence flow cytometry: Rapid
acidification of insulin to pH 6. J. Cell Biol. 98, 1757–1762.
Opladen, T., Ramaekers, V.T., Heimann, G., and Blau, N. (2006). Anal-
ysis of 5-methyltetrahydrofolate in serum of healthy children. Mol.
Genet. Metab. 87, 61–65.
Paulos, C.M., Reddy, J.A., Leamon, C.P., Turk, M.J., and Low, P.S.
(2004). Ligand binding and kinetics of folate receptor recycling
in vivo: Impact on receptor-mediated drug delivery. Mol. Pharmacol.
Sagne, C., Agulhon, C., Ravassard, P., Darmon, M., Hamon, M., El
Mestikawy, S., Gasnier, B., and Giros, B. (2001). Identification and
characterization of a lysosomal transporter for small neutral amino
acids. Proc. Natl. Acad. Sci. USA 98, 7206–7211.
Said, H.M. (2004). Recent advances in carrier-mediated intestinal
Selhub, J., and Rosenberg, I.H. (1981). Folate transport in isolated
brush border membrane vesicles from rat intestine. J. Biol. Chem.
Sharif, K.A., and Goldman, I.D. (2000). Rapid determination of mem-
brane transport parameters in adherent cells. Biotechniques 28,
Shayeghi, M., Latunde-Dada, G.O., Oakhill, J.S., Laftah, A.H., Takeu-
chi, K., Halliday, N., Khan, Y., Warley, A., McCann, F.E., Hider, R.C.,
et al. (2005). Identification of an intestinal heme transporter. Cell 122,
Stover, P.J. (2004). Physiology of folate and vitamin B12 in health and
disease. Nutr. Rev. 62, S3–S12.
Wang, Y., Zhao, R., and Goldman, I.D. (2004). Characterization of a
folate transporter in HeLa cells with a low pH optimum and high affinity
for pemetrexed distinct from the reduced folate carrier. Clin. Cancer
Res. 10, 6256–6264.
Wang, Y., Rajgopal, A., Goldman, I.D., and Zhao, R. (2005). Preserva-
tion of folate transport activity with a low-pH optimum in rat IEC-6
intestinal epithelial cell lines that lack reduced folate carrier function.
Am. J. Physiol. Cell Physiol. 288, C65–C71.
Wike-Hooley, J.L., Haveman, J., and Reinhold, H.S. (1984). The rele-
vance of tumour pH to the treatment of malignant disease. Radiother.
Oncol. 2, 343–366.
Zhao, R., Gao, F., Hanscom, M., and Goldman, I.D. (2004). A promi-
nent low-pH methotrexate transport activity in human solid tumor
cells: Contribution to the preservation of methotrexate pharmacologi-
calactivityinHeLa cells lacking thereduced folatecarrier.Clin.Cancer
Res. 10, 718–727.
Zhao, R., Chattopadhyay, S., Hanscom, M., and Goldman, I.D.
(2005a). Antifolate resistance in a HeLa cell line associated with
impaired transport independent of the reduced folate carrier. Clin.
Cancer Res. 10, 8735–8742.
Zhao, R., Hanscom, M., and Goldman, I.D. (2005b). The relationship
between folate transport activity at low pH and reduced folate carrier
function in human Huh7 hepatoma cells. Biochim. Biophys. Acta
928 Cell 127, 917–928, December 1, 2006 ª2006 Elsevier Inc.